Due to the size of the music collections of many users and particularly with the advent of subscription based music services providing unlimited access to millions of songs, recommendation technologies are emerging as an important enabler in assisting users to identify and navigate large databases of available music. Social networks provide an important environment for mining music recommendations. Recent studies have shown that peer pressure has as much to do with what users listen to as personal preferences. However, not all peers are equal. Celebrities, for example, display a disproportionate influence in setting trends in terms of what is popular. Applying this principle to music recommendations, there is a need for a system and method that provides media recommendations, such as music recommendations, based on the media presentations recently played by “heavy influencers” such as celebrities.